AI Is Cooking Up the Future, But Your Taste Buds Still Rule

The global AI food market is projected to explode from $10 billion in 2025 to over $50 billion by 2030. Yet in actual test kitchens, the story is more complicated than the hype suggests.

The Quiet AI Revolution Already Happening

McCormick has been using AI in flavor development for nearly a decade. The company behind Frank's RedHot, Cholula, and Old Bay says AI has cut development timelines by 20-25% on average by identifying promising flavor combinations and filtering which ideas deserve physical testing.

Unilever goes even further. Its AI systems can test thousands of recipes digitally in seconds, getting to viable concepts with fewer physical trials. The company's Knorr Fast & Flavourful Paste was developed in roughly half the usual time, while AI modeling for Hellmann's Easy-Out squeeze bottle saved months of physical lab work.

"Human creativity and judgment lead the way, and AI is a tool to help us amplify our impact," said Annemarie Elberse, head of ecosystems, digital and data for foods R&D at Unilever.

The Startup Challenge: Overpromise, Underdeliver?

A new breed of startups including Zucca, Journey Foods, and AKA Foods are marketing "virtual sensory" platforms that promise to digitally screen recipes and predict consumer preferences before physical prototypes exist. They're essentially promising what food giants say they're already doing.

But food scientist Brian Chau, who has tested these platforms, is skeptical. "I think all the AI companies coming out are, to some extent, overstating what they can do," he said. "They need to attract investors, they need to build datasets, and they need real industry partners before any of this really works at scale."

When Chau tested one platform, the results were disappointing: "The output was basically what you'd get from any general AI system. There wasn't much added value without proprietary data from real companies."

The Biology Problem

The biggest obstacle isn't computing power—it's human biology. Dr. Julien Delarue, a professor of sensory and consumer science at UC Davis, explains the fundamental limitation: "Trying to predict what people will perceive from a complex mixture of compounds—the answer is no."

The reason? Human sensory perception is inherently variable. People perceive the same chemical compounds very differently depending on genetics, culture, experience, and personal history. "There is no such thing as the average consumer," Delarue said. "Trying to predict what the 'average' person may perceive is probably a dead end."

Even IBM, which partnered with McCormick on early AI food projects and created Chef Watson, has moved on. An IBM spokesman confirmed the company is "not actively focused in this area anymore."

Why Humans Still Hold the Fork

David Sack, founder of AKA Foods, acknowledges the limitations: "It does not replace food scientists or sensory experts. Ultimately, humans define the goals, constraints, and success criteria. AI can reduce the number of tests needed, but it does not eliminate the need for real human tasting."

Jason Cohen, founder of Simulacra Data, puts it simply: "Consumers decide with their palate whether they like a product. We still start with real human sensory data. AI just helps us extrapolate insights faster and cheaper."

The real value, according to Delarue, lies in efficiency rather than creativity: "Designing food today is much more challenging than before. You need to make food that is healthy, sustainable, and affordable. AI gives us more power to handle that complexity."

The Investment Reality Check

While the $50 billion market projection sounds impressive, the path there remains unclear. Most current AI food platforms resemble large language models trained on existing recipes rather than systems capable of independently generating breakthrough products.

The winners will likely be companies that already control vast amounts of proprietary data—the McCormicks and Unilevers of the world—rather than startups trying to crack the code from the outside.